1. Field of the Invention
The present invention relates to pressure swing adsorption systems and processes for performing pressure swing adsorption.
2. Discussion of the Background
Pressure Swing Adsorption (PSA) is a technique used to fractionate mixtures of gases to provide at least one purified product gas and a raffinate byproduct mixture. PSA has been successfully used to separate hydrogen from other gases, oxygen and nitrogen from air, helium from natural gas, among others.
Early PSA systems generally used four adsorbent vessels operated in parallel. An example of this is U.S. Pat. No. 3,430,418 to Wagner. Later improvements to Wagner's process added an additional pressure equalization step while retaining four adsorbent beds (e.g., U.S. Pat. No. 3,564,816 to Batta) and subsequently added even more pressure equalization steps to seven or more beds in U.S. Pat. No. 3,986,849 to Fuderer et al. These increases in the number of pressure equalizations and the number of adsorbent vessels were implemented to increase the product recovery and the adsorbent productivity. Unfortunately, the increases in performance were accompanied by an increase in the number of valves required from thirty-one for the Wagner process to thirty-three for the Batta process to a minimum of forty-four for the Fuderer et al. process.
The performance of PSA cycles is commonly measured based upon several criteria. The first is product recovery at a given impurity level, the fraction of the product species in the total feed stream that is delivered as purified product. A second measure is the productivity of the adsorbent, which is related to the proportion of the PSA cycle during which product is delivered compared to the total length of the cycle. In order to maximize one or both of these parameters at fixed feed compositions, a number of approaches have been described in other systems.
Wagner describes the use of gas stored in the pressurized beds to repressurize one other vessel which had been purged, then to subsequently purge another vessel before the pressure in the first vessel was depleted. Batta subsequently describes that a second pressure equalization could be added to the first, and that this would improve recovery meaningfully. Batta retained the provision of purge gas in his cycle. Fuderer et al. extended this approach to a third pressure equalization, and taught that the purest gas withdrawn from a bed should always be the last gas admitted to any other bed being repressurized. Batta's four vessel cycle was constituted such that less pure gas was admitted to the vessel being pressurized than was truly desirable. Further, Fuderer et al.'s invention allowed for a higher adsorbent productivity than was achievable with previous cycles, because the fraction of time in the cycle allocated to adsorption was higher due to the details of the valve switching logic.
Although these methods facilitate excellent product recovery and adsorbent productivity, they do so at the expense of a high degree of complexity. Wagner's original process employed four vessels and thirty-one valves to facilitate one pressure equalization, and purging of one other vessel. Batta increased this total to thirty-three valves and four vessels for his cycle with two equalizations. Both of these four bed cycles produce gas from a given vessel twenty-five percent of the time. Batta also provided a five vessel system with forty-three valves to re-order the equalizations to provide the desired repressurization with gases increasing continuously in purity. This cycle produced from a given vessel only twenty percent of the time. Fuderer et al.'s most simple cycle providing three equalizations and a purging step required nine vessels and fifty-five valves. This cycle produced thirty-three percent of the time, a significant increase over the cycles of Batta and Wagner. Although these cycles progressed in the critical areas of recovery and productivity, they did so at the expense of much increased mechanical complexity. This increase in complexity is accompanied by increases in system volume, mass, assembly time, and capital cost. Furthermore, the large increase in the number of valves over time significantly reduces the reliability of the PSA system; as such PSA systems are single point of failure systems, which must be shut down even if one valve fails.
Efforts have been made to reduce complexity in order to address its attendant problems. U.S. Pat. No. 4,761,165 to Stöcker implemented the process of Wagner using four vessels and eighteen valves, of which four could be proportionally-controlled valves. U.S. Pat. No. 6,146,450 to Duhayer et al. describes a means for reducing complexity by arranging pipe fittings optimally, although this approach does not materially alter the PSA cycle in terms of valve or vessel count. Much more dramatic mechanical simplification has been described in an application entitled METHOD AND APPARATUS FOR PRESSURE SWING ADSORPTION by inventor Franklin D. Lomax, Jr. et al. filed concurrently herewith that is based upon Provisional Application No. 60/370,702, entitled IMPROVED METHOD AND APPARATUS FOR PRESSURE SWING ADSORPTION, which was filed on Apr. 9, 2002.
Although Stöcker's process does attain meaningful simplification in the number of valves used relative to the process of Wagner, he does not reveal means for accomplishing a second or third pressure equalization to increase recovery, nor does he describe an increase in adsorbent productivity. The mechanical simplifications of both Duhayer et al. and Lomax et al. offer advantages for virtually any PSA cycle, and do not specifically teach any inherent reduction in valve count or changes in cycles to achieve increased recovery or adsorbent productivity without the attendant increases in valve and vessel usage taught by Batta, Fuderer et al. and others. Furthermore, none of these improvements fundamentally alter the reliability of the PSA cycles, as the basic cycles remain single point of failure systems.